Lecture 05Chapter 4 Capital Market Expectations Strategy/Economics team 2Portfolio Management ProcessPLANNING Capital Market Expectations E(r)/PLANNING Investor Objective & ConstraintFEEDBACK Performance/Monitoring Performance measures Sp Attribution analysis REBALANCEPLANNING Strategic Asset Allocation Efficient Frontier Based on objective/constraints, max return on risk adjusted basisEXECUTION Tactical Asset Allocation Security Analysis Transaction Costs3FOCUS of LECTURE: 1. Framework & Challenges2. Tools for Formulating Capital Market Expectations3. Economic Analysis4General DefinitionsCapital Market Expectations of risk & returns Long term forecasts strategic asset allocation (SAA) Short term forecasts tactical asset allocation (TAA) Macro expectations forecast of asset classes & sectors Micro expectations forecasts of individual assetsCapital Market Expectations is beta research systematic risk & systematic returnvsalpha research (non systematic) excess risk-adjusted returns 5Framework (something you can do)Define asset classes, time horizon & tax status Compile historical return, risk of asset classes Identify factors that drive of return & risk of asset classSelect model used to forecast return & risk of asset classIdentify best data source for inputs into modelMonitor performance & reviseInterpret forecasts from model Apply professional judgment based on experienceFormulate capital market expectations6Problems encountered in forecastingNumber of factors may impact on Capital Market Expectations Limitations of economic data (time lag, revisions, change method) Data measurement errors & biases Limitations of historical estimates Failure to account for conditioning information Ex post risk (after fact) can be biased measure of ex ante (before fact) risk Bias in analysts methods Misinterpretation of correlations Psychological traps Model uncertaintyIssues you need to consider in your research project7Problems encountered in forecastingData measurement errors & biases Transcription error (ignore outliers) Survivorship bias (only include data that survived at end period) Use of appraisal (smoothed) data possible to adjust by increasing volatilityLimitations of historical estimates Regime changes resulting in nonstationary data tradeoff between amount of data vs stationary data (mean & standard deviation do not change with time)Failure to account for conditioning information Identification of relationship which ignores 2nd conditioning factor rather than evaluate 1 long time series data should be separated into 2 ranges (economic expansion vs recession)Bias in analysts methods Data mining identification of apparently statically significant relationship but is random event Time-period bias8Psychological Traps (Behavourial)Reasons why analyst are unable to make accurate & unbiased forecasts Behavioural Finance Anchoring trap place too much weight on 1st information received Status quo trap perpetuate recent observation (ie predict NO change from recent past) Confirming evidence trap place greater weight on information that supports preferred point of view rather than consider evidence that contradicts point of view Overconfidence trap individuals overestimate accuracy of their forecasts & have too narrow range of possibilities Prudence trap tendency to Temper forecasts so they do not appear to be too extreme Be overly cautious in forecasting Recallability trap tendency to be overly influenced by events or individuals that have left strong impression on persons memory 9ToolsTypes of tools to develop Capital Market Expectations Statistical Methods Descriptive statistics characterizes historical data set Inferential statistics making forecasts utilizing historical data Discounted Cash flow Models Equity Markets Fixed Income Markets Risk Premium Approach (build-up approach)10Tools: Statistical MethodsHistorical Statistical Approach Determine historical mean return (R), standard deviation (), correlation () of returns of various asset classes Underlying assumption is data is stationary & historical probability of returns can be used for future returns Historical average return vs Historical geometric returnQUESTION: What is mean & standard deviation of Canadian Equity, & Canadian Fixed income? every one should know thisShrinkage Estimators Weighted average of historical combined with estimate (forecast) w = weight applied to historical data Cov = covarianceShrinkage Estimator of Cov=W CovHistory+(1 w) CovAnalyst fprecast11Tools: Statistical MethodsInferential statistics Multi factor models Forecasting variable (return & risk of asset class or individual security) in terms of set of risk factors Consider 2 factor model Ri = actual return of asset i return of Canadian TSX ai = constant Fk = return due to factor k bik = se